1. Field of the Invention
This invention relates to a networked material sorting system for sortation of units of material.
2. Description of Related Art
Industrial networks for high-speed handling of material are used in applications such as mail handling, parcel distribution, warehouse distribution and airport baggage distribution. Such industrial networks rely upon accurate and time efficient handling using transport media such as conveyor belts for moving the material through the handling process. Sensors and actuators are variously used to generate a signal and execute an appropriate handling action based upon such signal.
Prior industrial networks use a time segmented approach to organize and rule the media access of the industrial network. Such industrial networks utilize technology known as master-slave principle, token passing or time multiplex. The result of such technology is that prior industrial networks exchange information in cycles. A signal from a sensor or actuator is only reported at a particular interval within the cycle. Therefore, if a sensor or actuator changes condition, the new condition cannot be reported until an entire cycle is completed. This results in variation of media access time for input signals from actuators and sensors.
Another common problem with prior industrial networks is that multiple actuators and/or sensors do not get network data at exactly the same time. Instead, a hierarchy within the cycle is executed and actuators and/or sensors only receive network data on an xe2x80x9cas-neededxe2x80x9d basis. Therefore, the accuracy of the industrial network is dependent upon how quickly all of the actuators and/or sensors can be updated.
Prior industrial networks typically contain control logic engines that are based on micro controllers that execute an execution code in a serial manner. The basic routine common in prior industrial logic controllers is to read inputs, calculate the outputs based on the latest input information and set the outputs on the end of the execution, or logic, cycle. After this, the next cycle starts with reading the inputs. Even modern high-speed controllers incorporating intermediate sub-cycles operate in this manner. Such modern high-speed controllers have main cycles which can be interrupted by smaller, faster sub-cycles that execute in a similar manner.
By way of further explanation, regular cyclical control systems such as a PLC (Programmable Logic Controller) execute their program in an endless loop which has 3 stages.
1) Read inputs
2) Calculate outputs (Executing program)
3) Write outputs
Then the controller loops back to stage 1. This leads to the following cycle pattern: . . . 1)2)3) 1)2)3) 1)2)3) 1)2)3) . . . .
The PLC tries to execute its program as fast as possible, therefore the cycle time (time required to execute 1)2)3)) depends on the program size. It is also true that this cycle time is not necessarily constant due to the fact that the program can have several different execution paths depending the programmed internal decisions driven by the particular input pattern at a given time.
In an earlier generation of sortation control systems, the inputs and the outputs were wired up directly to the main PLC, sometimes also called a micro controller or controller. In other words the micro controller read the inputs directly from the sensors/switches on the sortation line and activated the actuators for moving packages, such as valves, motor starters, etc in a direct fashion.
Over time the industry began to move to adapted networks to reduce wiring. By moving from hardwired solutions to network solutions the industry replaced a parallel data exchange system with a serial system. In practice this meant that the input/output signal conditioning card in the main controller got replaced with a network control layer having its own card of electronics including its own network controller, i.e. micro controller or PLC. The main, or sort logic, controller still reads the 1) inputs 2) executes its program and 3) writes the outputs However, the main controller does not have direct access to the inputs and outputs of the sensors anymore, but must communicate to the sortation line through the network controller, or control layer. The main controller thus reads from and writes to memory locations of the network card.
The network controller is responsible to deliver the output signals of the main controller to the different network devices, i.e. those devices on the sortation line. The network controller is also responsible to read the inputs of the network devices and place their data into the memory where the statuses can be read by the main controller. The network controller works independent of, i.e. not synchronized with, the main controller (PLC). In other words the network controller does not care if the main controller reads inputs or sets outputs. The network controller is continuously reading and writing data to and from network stations data in a serial matter. There are several different ways how a network controller in detail can realize this data exchange. In general the network controller runs in a loop as well, generally described as:
A) Read output data from memory
B1) Deliver output data to station 1 (network device)
B2) Deliver output data to station 2 (network device)
B3) Deliver output data to station 3 (network device)
Bx)
C1) Read input data from station 1 (network device)
C2) Read input data from station 2 (network device)
Cx)
D) Write input data into memory (network device) After this process the network controller loops back and starts new cycle, resulting in the pattern:
. . . A)B)C)D) A)B)C)D) A)B)C)D) A)B)C)D) . . .
Dependant on the network speed, the utilized protocol, and the number of inputs and outputs, the network cycle time can vary.
FIG. 7 shows the cycles of the main controller and the network controller in a nonsynchronized arrangement. In such a construction the delivery times for outputs (the time between main controller writing to memory and reception of signal by actuator, i.e. a network device) are not constant and can vary greatly between the time of two outputs in a single logic cycle, i.e. the difference in length between To1 and To2, and the difference in execution time between two cycles, i.e. the difference between To1 and To1xe2x80x2 and To2 and To2xe2x80x2. Due to these different cycle and delivery times, as a practical matter the industry uses the following rules: the Network cycle has to be half (or faster) the time of the Controller cycle, which leads to a worst case overall system response time of 2x controller cycle time+2x network cycle time.
What is therefore needed is a way of dramatically improving the system response time and generating a situation where the output delivery time delays are constant, in order to effectively control a sorting system over a serial network.
The present invention accomplishes the above desired tasks by a technique sometimes called xe2x80x9csynchronized logic solve and output deliveryxe2x80x9d combined with a multicasting of the output signals to the network devices. xe2x80x9cSynchronized logic solve and output deliveryxe2x80x9d refers to the fact that the control solution output (signal) derived by the main, or sort logic, controller logic cycle is written to the network control layer and read by the network controller in a synchronized fashion, and that control solution output (signal) is delivered without delay to the network devices, so that no irregularity of cycle time is experienced in operating the network devices on the sortation line according to the present invention.
Referencing FIGS. 7 and 8, the difference between the known systems of FIG. 7 and the present invention of FIG. 8 are illustrated graphically. In the present invention, the output signals of the sort logic controller are delivered through the network control layer without delay by a multicast message, which means that instead sending single messages or outputs to each network station, or device, the network controller sends out a single message containing all the output data for the entire system, i.e. every station or device. All network stations thus receive this message at the same time thereby eliminating any time differences between To1 and To2, as seen in FIG. 8.
Further, according to the technique of synchronized logic solve and output delivery, the network controller is waiting for a signal, such as an interrupt, from the main controller directing that all outputs are to be written to the memory. While the signal is designated as an interrupt, the artisan of ordinary skill in the field of network-controlled high-speed sortation devices will appreciate that there are several different ways to implement this technique. At the interrupt the network controller delivers the outputs 48 to the network devices and continues the network cycle by reading all inputs and then waiting again for the synchronization signal, as seen in FIG. 8.
It is one object of this invention to provide a material sorting system in which a plurality of actuators, or other network devices receive a multicast message from a network logic controller.
It is another object of this invention to provide a material sorting system wherein a transport media has segments and an adjacent segment sensor for measuring the speed of the transport media.
It is another object of this invention to provide a material sorting system to multicast an output signal received from at least one of a segment sensor, an identification sensor and a length sensor to a plurality of actuators.
It is another object of this invention to provide a material sorting system having actuators capable of accurately diverting units of material at high speeds.
Several attempts to apply serial communication networks to material handling and sorting applications have been unsuccessful because of resolution and accuracy problems owing to the above discussed deficiencies.
A sort system for material handling according to one preferred embodiment of this invention comprises an In-Feed Section in which units of material are fed to a main sorting line; a Singulation Section where the units of material are identified, characterized and/or distinguished and placed on a transport media with a predetermined interval between each unit of material; and a Handling Section where the units of material are diverted to a correct divert line. The sort logic, or main, controller calculates the necessary output signals for a plurality of actuators on an event driven basis based on information provided by at least one of a length sensor, a segment sensor and an identification sensor that are each positioned with respect to the main sorting line.
The transport media, such as a conveyor belt, is segmented in physical segments. Although the logical segmentation of the transport media is paramount, physical segments of the transport media are preferably of equal size. The sort logic controller represents those segments as a logical cell in a shift register. An identification sensor for identifying and/or distinguishing units of material and a plurality of actuators are preferably positioned along fixed positions with respect to the transport media. The actuators are continuously provided with the shift register status, which indicates the physical location of the respective divert lines and/or the actuators within the material sorting system. The actuator is activated when the actuator identifies a match between the physical location of the actuator and the respective xe2x80x9ctickxe2x80x9d in the logical shift register.
Two principal changes are required to overcome the demands of the industrial network according to this invention. First, the time chain must be synchronized. This requires that the sort logic controller and the network controller of sensors and actuators do not have two independent rotating cycles. An event-driven architecture is required that allows transport, evaluation and generation of information based on events within the material sorting system. Second, the notification and activation of the actuators must be multicast so that all actuators which have to be fired in one system cycle react exactly at the same time. Additionally, in one preferred embodiment of this invention, each actuator should calculate an exact activation point using the actual speed of the transport media, the actual position of the unit of material and the length of the unit of material.
The actual speed of the transport media can be calculated by measuring the elapsed time between segments using a segment sensor. The measurement of transport media speed may be simplified in the preferred embodiment of this invention wherein the segments of transport media are each the same size. The segment sensor preferably multicasts the passing of a change from segment to segment so that all actuators receive notice of the change. The segment sensor can multicast either the actual speed of the media, the time between the segment changes or the event of the segment change. It is important that such multicast has the highest network priority